Mike Rivers wrote:

[ ... ] X-Y recording is mostly off axis (where they don't measure
frequency response when "matching" mics) matching wouldn't really
help that either.

Actually, the one manufacturer I know that does still do pair matching of
microphones (or larger sets than two if needed) does consider sensitivity
and both on- and off-axis response. It's for exactly the reason you gave:
In coincident stereo recording you don't aim either microphone's main axis
directly forward, unless it's the "M" microphone in an M/S pair of
figure-8s--which is not a very often-used recording technique.

But just between you and me and 10,000 lurkers, I suspect that if certain
influential producers and engineers were to ask Neumann Berlin nicely
to hand-select a pair of microphones, Neumann would at least check the
factory curves to make sure that the proposed pair was not a bad match.
They're polite, reasonable, service-oriented people; they're not in
business to offend their customers. Just because something isn't on the
price list doesn't mean that they won't consider it if you ask nicely.

AKG used to be the same way and back in the 1970s I was offered special
pair matching if I wanted it. But I've lost touch with the people I knew
there back then, so I don't know if that goes on any more now that they're
A Harman Company and all.